Muleta Environ Syst Res (2021) 10:24 https://doi.org/10.1186/s40068-021-00225-5

RESEARCH Open Access Climate change scenario analysis for Baro‑ river basin, Southwestern Teressa Negassa Muleta*

Abstract Background: Several water resources projects are under planning and implementation in the Baro-Akobo basin. Cur- rently, the planning and management of these projects is relied on historical data. So far, hardly any study has addressed water resources management and adaptation measures in the face of changing water balances due to climate change in the basin. The main bottleneck to this has been lack of future climate change scenario base data over the basin. The current study is aimed at developing future climate change scenario for the basin. To this end, Regional Climate Model (RCM) downscaled data for A1B emission scenario was employed and bias corrected at basin level using observed data. Future climate change scenario was developed using the bias corrected RCM output data with the basic objective of producing baseline data for sustainable water resources development and management in the basin. Result: The projected future climate shows an increasing trend for both maximum and minimum temperatures; however, for the case of precipitation it does not manifest a systematic increasing or decreasing trend in the next century. The projected mean annual temperature increases from the baseline period by an amount of 1 °C and 3.5 °C respectively, in 2040s and 2090s. Similarly, evapotranspiration has been found to increase to an extent of 25% over the basin. The precipitation is predicted to experience a mean annual decrease of 1.8% in 2040s and an increase of 1.8% in 2090s over the basin for the A1B emission scenario. Conclusion: The study resulted in a considerable future change in climatic variables (temperature, precipitation, and evapotranspiration) on the monthly and seasonal basis. These have an implication on hydrologic extremes-drought and fooding, and demands dynamic water resources management. Hence the study gives a valuable base informa- tion for water resources planning and managers, particularly for modeling reservoir infow-climate change relations, to adapt reservoir operation rules to the real-time changing climate. Keywords: A1B emission scenario, Baro-Akobo River Basin, Bias Correction, Climate change scenario, General circulation model, Regional climate model

Background and infrastructures. Tis assumption is nowadays threat- Te planning and management of basin scale water ened by climate change and the notion that both climate resources systems has historically relied on assuming and hydrology will evolve in the future. Water resource stationary hydrologic conditions. Tis approach assumes planning based on the concept of a stationary climate is that past hydrologic conditions are sufcient to guide the increasingly considered inadequate for sustainable water future operation and planning of water resources systems resources management (Xueping et al. 2018; Mohamed 2015; Luca et al. 2020). According to the Series Assess- *Correspondence: [email protected] ment Reports of the Intergovernmental Panel on Climate Department of Sanitary and Environmental Engineering, Budapest Change (IPCC 2007) and many other studies, warming of University of Technology and Economics, Budapest, Hungary

© The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​ mmons.​org/​licen​ses/​by/4.​0/. Muleta Environ Syst Res (2021) 10:24 Page 2 of 15

the climate system is unequivocal, as is now evident from downscaling approach in which a fne, at a much smaller observations of increases in global average air and ocean space-scale (e.g., 0.5° by 0.5°), computational grid over a temperatures, widespread melting of snow and ice, and limited domain is nested within the coarse grid of a GCM rising global average sea level (Myo and Zin 2020; Rah- (Wilby and Wigley 1997; Hewitson and Crane 1996; man 2018; Safeh et al. 2020; Jose et al. 2016). Cubash et al. 1996; Mearns et al. 1999). Adaptation is recognized as a critical response to the impacts of climate change. It can reduce pre- Climate and emission scenarios sent and future losses from climate variability and A climate scenario is a plausible representation of future enhance climate-resilient futures (Conde et al. 2011; climate that has been constructed for explicit use in inves- Cradock-Henr et al. 2018). It is a process that needs tigating the potential impacts of anthropogenic climate to be incorporated in the overall development plan- change (IPCC 2001). Te climate change scenarios should ning, including the design and implementation of be assessed according to consistency with global projec- water resources development. Over the last few tions, physical plausibility, applicability in impact assess- years, the literature on adaptation to climate change ments and representativeness (Hulme and Viner 1998). has expanded considerably worldwide: (Schneider Climate change scenarios are tightly dependent on the et al. 2000; Brekke et al. 2009; Li et al. 2010; Raje and emission scenarios. Future greenhouse gas (GHG) emis- Mujumdar 2010; Jones 1999). However, hardly few sions are the product of very complex dynamic systems, studies have addressed water management and adap- determined by driving forces such as demographic devel- tation measures in the face of changing water balances opment, socio-economic development, and technological due to climate change in Ethiopia. change (Luca et al. 2020). Scenarios are alternative images of how the future might unfold and are an appropriate tool Climate models and downscaling that assist in climate change analysis. According to IPCC General circulation models (GCMs) are tools designed Working Group III Special Report on Emission Scenario to simulate time series of climate variables globally, (SRES), four diferent narrative storylines were developed: accounting for efects of greenhouse gases in the atmos- (1) Te A1 storyline that describes a future world of very phere. Tey attempt to represent the physical processes in rapid economic growth, global population that peaks in the atmosphere, ocean, cryosphere, and land surface. Tey mid-century and declines thereafter and the rapid intro- are currently the most credible tools available for simulat- duction of new and more efcient technologies; (2) Te ing the response of the global climate system to increasing A2 storyline that describes a very heterogeneous world; greenhouse gas concentrations, and to provide estimates of (3) Te B1 storyline that describes a convergent world with climate variables (e.g., air temperature, precipitation, wind the same global population as in A1; and (4) Te B2 sto- speed, pressure, etc.) on a global scale. GCMs demonstrate ryline that describes a world in which the emphasis is on a signifcant skill at the continental and hemispheric spatial local solutions to economic, social, and environmental scales and incorporate a large proportion of the complexity sustainability. of the global system; they are, however, inherently unable to In the current study, the A1B storyline was chosen as it represent local sub grid-scale features and dynamics, which represents a balanced (moderate) future scenario in terms are of interest to a hydrologist (Myo and Zin 2020; Conde of the alternative energy, fossil intensive, socioeconomic, et al. 2011; Feyissa et al. 2018). and developmental trajectories. Te A1 scenario family Poor performances of GCMs at local and regional scales develops into three groups that describe alternative direc- have led to the development of limited area models (Dick- tions of technological change in the energy system as fos- inson et al. 1989; Dinckison and Rougher 1986; Anthes sil intensive (A1FI), non-fossil energy sources (A1T), or a et al. 1985). Tese techniques known as regionalization or balance across all sources (A1B). Even though it is true that downscaling are adopted in the literature to describe a set the occurrence of a single storyline is highly uncertain, the of techniques that relate the local and regional scale climate A1B scenario is adopted in this paper to have single-val- variables to the large scale atmospheric and oceanic forc- ued climate variable predictions instead of a ranged out- ing variables. Diferent regionalization techniques are avail- come from using ensemble of extreme trajectories. Tese able in the literatures so far: Spatial downscaling, Statistical projected climatic variables can, thus, directly be used for Downscaling, and Dynamic Downscaling (Barrow et al. climate change adaptation purposes like modeling climate 1996; Conway and Hulme 1996; Smith and Pitts 1997). change-runof relationships. For the current study, the dynamic downscaling method So far diferent studies have been conducted on the cli- (regional climate model version 3.1) was employed by mate change scenario development across the world using International Water Management Institute, Ethiopia, to diferent models and emission scenarios. Xueping et al. downscale the GCM output data at regional level. It is a (2018) with their study on Biliu River, China, has shown Muleta Environ Syst Res (2021) 10:24 Page 3 of 15

that climate emission scenarios are closely associated with in the north) has projected an increasing scenario both in temperature but not so closely associated with precipitation rainfall and temperature into the future (Roth et al. 2018). and runof; Conde et al. (2011) have concluded from his Furthermore, the study by Hany et al. (2016) has predicted study that confdence on future climate change estimations a considerably increasing precipitation into future scenario is greater for temperature than for precipitation; According on (Baro-Akobo-Sobat basin). Below is an to the study by Poonia and Rao (2018), rainfall trend during overview of selected studies on climate change scenarios the last 100 years revealed that the summer monsoon rain- to show anomalies in rainfall change scenarios for com- fall has increased marginally (< 10%) in the southern and parison (Table 1). eastern parts of the Tar Desert but has already declined by 10–15% in its northwestern part India. A study in Angola Statement of the problem by Carvalho et al. (2017) has also shown the precipitation Despite the numerous climate change scenario stud- projections to be highly variable across the region with the ies mentioned above, it is clearly visible that there exists southern region experiencing a stronger decrease in pre- inconsistency in the rainfall change scenario among dif- cipitation and (Costa et al. 2019), with their study on Recife, ferent models (Table 1); spatial and temporal variation in Brazil, have emphasized the considerable decrease of rain- the result from a single projection scenario (Carvalho et al. fall in the rainy season from March to August as a contrast 2017; Poonia and Rao 2018); and even seasonal variation to an increase of rainfall in the dry months, from Septem- from the same study (Costa et al. 2019) is very visible in the ber to February. rainfall change scenario results. Tis is mainly attributed to Furthermore, NAPA (2007) has revealed that in Ethi- the sensitivity of the rainfall variable to local efects of par- opia climate variability and change in the country is mainly ticularly orographic, coastal and vegetation efects in gen- manifested through the variability of rainfall and temper- eral. Particular to the current study, the Ethiopian climate ature. Te trend analysis of annual variables for (1951– patterns show large regional diferences and is also charac- 2006) shows that rainfall remained roughly constant when terized by a history of climate extremes, such as drought averaged over the whole country while temperature shows and food (NAPA 2007). Baro-Akobo basin exhibits dis- an increasing trend all over the country. Alemayehu et al. tinctive bimodal rainfall pattern diferent from most parts (2016) with his study of historical climate analysis on Baro- of the country. Tere are also numerous water resources Akobo River Basin (1952–2009) has shown an increasing development projects under planning and implementation trend for temperature and a declining trend in case of rain- in the basin (Tams, Birbir A & Birbir R, Baro-1 and Baro-2, fall. On contrary, the study on (bordering basin Geba-A and Geba-R, and Genji hydropower projects; and

Table 1 Overview of previous studies on climate change scenarios Study area Emission scenario Study period Major fndings References Base period Future ΔT (°C) ΔP (%)

Biliu River, China Ensemble of scenarios 1980–2004 2016–2040 0.4–1.5 12.2 to 21.8 Xueping et al. (2018) + − + 2041–2065 0.68–3.4 4.1 to 2.2 + − + Angola RCP4.5 1958–1987 2011–2040 1.3 2.9 Carvalho et al. + − (2017) 2041–2070 2.1 1.3 + − 2071–2100 2.6 2.6 + − Dry Zone in Myanmar RCP4.5 1981–2005 2050s 1.5–2.2 10 Myo and Zin (2020) + + 2080s 2.0–2.6 8 + + Recife city, Brazil RCP4.5 1979–2000 2021–2050 1.0 5.2 Costa et al. (2019) + − 2051–2080 2.0 15.4 + − Over Ethiopia A1B 1961–1990 2030 0.9–1.1 1.4–4.5 NAPA (2007) + + 2050 1.7–2.1 3.1–8.4 + + 2080 2.7–3.4 5.1–13.8 + + Over White Nile Basin A2 Existing condition 2010–2039 1 1.16 Hany et al. (2016) + − 2040–2069 2.5 13.58 + − 2070–2099 3.5 14.77 + − Over Blue Nile A1B 1979–2013 2046–2064 2.0–2.7 17.7–46 Roth et al. (2018) + + 2081–2099 2.7–3.7 27–48 + + Muleta Environ Syst Res (2021) 10:24 Page 4 of 15

Itang and Gilo-2 irrigation projects) that will have environ- with in their downstream sections and merge mental impacts on the local and downstream Nile coun- to form the , which is a major tributary of tries—Sudan and Egypt (Tahani et al. 2013). White Nile. Neighboring river basins in Ethiopia are Terefore, it is very essential to develop basin level climate the Abay river basin (Blue Nile) in the north and the change scenario to address water resources management Omo- basin in the south-east. The river issues related to climate change in the basin. Unfortunately, basin has lowest elevation of about 390 m and high- there still lacks a full-fedged climate change scenario devel- est elevation of about 3244 m. The total mean annual opment at Baro-Akobo basin level. Furthermore, previ- flow from the river basin is estimated to be 23.6 BCM. ous studies used to apply RCM outputs directly for climate As a result of regular flooding, the lowland areas are change scenario development whereas the current study has mainly used as pastures for grazing and no major put an efort to treat the biases in the regionally downscaled water resources development has taken place to-date. time serious data due to local factors (orographic, coastal There are Extensive Water Resources Development and vegetation efects) using 30-years observed data. Projects under planning and implementation in the basin. These include hydropower projects—Tams, Objective of the study Birbir A, Birbir R, Baro-1, Baro-2, Geba-A, Geba-R, Climate change adaptation is fundamental to safeguarding and Genji hydropower projects and Irrigation projects vulnerable communities, ecosystems, and relevant climate- including Itang and Gilo-2 irrigation schemes (Fig. 1). sensitive sectors from the impacts of climate change. Te key to successful adaptation is to anticipate how the climate Data collection and processing will evolve in the future (Santosh et al. 2019). To understand Observation data the impacts of climate change on diferent sectors, the his- Moreover, spatial data such as Digital Elevation Model torical trends and future climate scenarios of diferent cli- (DEM) of 90 m*90 m, and Ethio-River basin shape matic parameters need to be assessed (Cradock-Henry et al. fles were collected from Ministry of Water Resource 2018). (MoWR), Ethiopia. Tis data is applied for location map Terefore, the main objective of this study is to develop development and areal data computations, both for future climate change scenario in the mid-term and long- observed and RCM, from point data. Diferent data pro- term future periods using the A1B emission scenario. Te cessing techniques—data screening, trend test, stationery focus is to develop bias correction factors for RCM down- test, F-& t-tests, test for relative consistency and homo- scaled data and projection of future climate change in terms geneity are all applied to ensure observation data quality. of temperature, precipitation, and evapotranspiration for the basin using the A1B emission scenario. Te developed Mean‑Areal Precipitation (MAP) depth computations scenario provides a basis to design adaptation pathways Tere are many ways of deriving the areal precipitation such as modeling the relation between reservoir infow over a catchment from rain gauge measurements includ- and climatic variables based on the likely climate change ing: Arithmetic Mean, Tiessen Polygon, Isohyetal, grid trajectories. Point, Percent Normal, Hypsometric, etc. (Chow et al. 1988). Choice of methods requires judgment in consideration of Materials and methods quality and nature of the data, and the importance, use and Description of the study area required precision of the result. Accordingly, the Tiessen The Baro-Akobo River Basin lies in the South-West- Polygon method was applied to determine the areal aver- ern part of Ethiopia between latitudes 5° and 10° age precipitation over the basin. In this method, weights are North and longitudes 33° and 36° East. In the west given to all the measuring gauges based on their areal cov- the basin boundary forms an international boundary erage of the watershed, thus eliminating the discrepancies in with Sudan. The basin covers parts of the Benshangul- their spacing over the basin. HEC-GeoHMS utility of Arc- Gumuz, Gambella, Oromia and SNNP administrative GIS is employed for construction of Tiessen Polygons and regions. It is the second largest subbasin in the Eastern assigning gage weight. Similar procedure is followed to cal- Nile basin. The Eastern Nile Basin consists four sub- culate other basin areal meteorological data (e.g., ETo) and basins: the Baro-Akobo-Sobat (White Nile) sub-basin areal− data of RCM GCPs meteorological data.Te areal rain- in the west, the Abbay (Blue Nile) sub-basin in the fall R is given by: north, the Tekeze-Atbara sub-basin on the east and the n n aiRi Main Nile basin from Khartoum to the Nile delta. R = = w R A i i (1) With a total drainage area of about 76,000 sq km, i=1 i=1 the basin ranks number eight of the 12 major river basins in Ethiopia. Both Baro and Akobo rivers border Muleta Environ Syst Res (2021) 10:24 Page 5 of 15

Fig. 1 Location map of Baro-Aakobo river basin relative to Ethiopian river basins

0.408 R G 900 u e e where ­Ri is the rainfall measurement at n rain gauges; a­ i �( n − ) + γ T+273 2( s − a) ETo = is polygon area; w­ i is gauge areal weight and A is the total � + γ (1 + 0.34u2) (2) area of the catchment. −1 where ETo = reference evapotranspiration [mm day­ ], Potential evapotranspiration −2 −1 ­Rn = net radiation at the crop surface [MJ ­m day­ ], Many empirical or semi-empirical equations have been −2 −1 G = soil heat fux density [MJ m­ day­ ], T = mean developed for assessing reference/potential evapotranspi- daily air temperature at 2 m height [°C], ­u2 = wind speed ration from meteorological data. Numerous researchers −1 at 2 m height [m s­ ], ­es = saturation vapour pressure have analyzed the performance of the various calculation [kPa], ­ea = actual vapour pressure [kPa], e­ s-ea = saturation methods for diferent locations. As a result of an Expert vapour pressure defcit [kPa], ∆ = slope vapour pressure Consultation held in May 1990, the FAO Penman–Mon- −1 −1 curve [kPa °C ], γ = psychrometric constant [kPa °C ]. teith method is now recommended as the standard method for the defnition and computation of the poten- Baseline climate tial evapotranspiration when the standard meteorological Baseline climate information is important to characterize variables including air temperature, relative humidity and the prevailing conditions and its thorough analysis is valu- sunshine hours are available (Allen et al. 1998). able to examine the possible impacts of climate change on In this study, the Potential evapotranspiration is calcu- a particular exposure unit. It can also be used as a refer- lated by ETO calculator software that uses the relatively ence with which the results of any climate change studies accurate and consistent performance of the Penman– can be compared. Te choice of baseline period has often Monteith approach in both arid and humid climates. ETo been governed by availability of the required climate data. for the basin is determined based on the four basic cli- According to World Meteorological Organization (WMO), matic data—temperature ­(Tmax, ­Tmean and T­ min), mean the baseline period also called reference period gener- relative humidity, wind speed at 2 m above soil surface ally corresponds to the current 30 years normal period. A and the actual sunshine duration. 30-year (1989–2018) period is used by this study to defne Te FAO Penman–Monteith equation (Allen et al. the average climate of the basin, and scenarios of climate 1998)) is given by: change are also generally based on 30-years’ means. Muleta Environ Syst Res (2021) 10:24 Page 6 of 15

RCM data (Robert and Colin 2007). Bias correction is usually needed Regional grid climate data that has been dynamically as climate models often provide biased representations downscaled by RCM Version 3.1 from general Circu- of observed times series due to systematic model errors lation model (GCM) for A1B emission scenario was caused by imperfect conceptualization, discretization, and obtained from International Water Management Institute spatial averaging within grid cells. Typical biases are the (IWMI). Tese are gridded meteorological time-series occurrence of too many wet days with low-intensity rain data of precipitation; maximum, mean, and minimum or incorrect estimation of extreme temperature in RCM temperature; wind speed; relative humidity; and poten- simulations (Ines and Hansen 2006). A bias in RCM-sim- tial evapotranspiration at a grid center point (GCP) spa- ulated variables can lead to unrealistic hydrological simu- tial resolution of 0.5° by 0.5° (i.e., 55 km by 55 km). GCPs lations of river runof. Tus, application of bias-correction that can have possible efect and fall in or nearby the methods is recommended (Wilby et al. 2000). basin are selected for the study. Te data cover 3-periods: Te term ‘bias correction’ describes the process of scal- base (control) period (2001–2010), mid-term forecast ing climate model output to account for systematic errors (2040s i.e., 2041–2050) and long-term forecast (2090s i.e., in the climate models. Te basic principle is that biases 2091–2100) meteorological time-series RCM data. Tese between simulated climate time series and observations data are bias corrected based on bias correction factors are identifed and then used to correct both control and developed between means of observed base period data scenario runs. Te main assumption is that the same bias (1989–2018) and RCM base period data (2001–2010) correction applies to control and scenario conditions. before making use of them for mid-term and long-term Several techniques are available to create an interface for scenario development. translating RCM output variables to hydrological mod- els. For instance, precipitation and temperature can be Bias correction for RCM output bias corrected by applying one of the following methods: Tis method is used to ‘adjust’ the mean, variance and/or Precipitation threshold, Scaling approach, linear trans- distribution of RCM climate data. All models have inad- formation, Power transformation, Distribution transfer, equacies due to such factors as resolution, difering inter- Precipitation model and Empirical correction methods nal dynamics, and model parameterizations, so diferent (Fig. 2). climate models can respond diferently to the same inputs

Fig. 2 Conceptual framework for bias correction Muleta Environ Syst Res (2021) 10:24 Page 7 of 15

In this work a Power Transformation method of bias “Goal Seek” iteration technique of MS Excel so that the correction is applied for precipitation whereas linear statistics (mean and standard deviation) of the observed transformation method is used to correct temperature. In data corresponds with that of the corrected base period each case of bias correction attempt is made to match the RCM data. most important statistics (coefcient of variation, mean and standard deviation) on a scale of 30 days. Result Bias correction of RCM output climatic variables Precipitation bias correction: power transformation Bias correction factors—a and b method As thoroughly explained under methods section above, Te method is selected based on its relative simplicity bias correction factors a and b were developed by match- and application result as used by (Leander and Buishand ing the statistical parameters (mean, coefcient of 2007) for a Meuse basin study. Tey found that adjusting variation and standard deviation) of the observed and both the biases in the mean and variability, the method downscaled base period precipitation and temperature leads to a better reproduction of observed extreme daily data. Values of a and b were determined for each month and multi-day precipitation amounts than the commonly block of a year to minimize the temporal variation impact used linear scaling correction. on the bias correction and these values were applied to In this nonlinear correction each daily precipitation correct each corresponding month of the mid-term amount P is transformed to a corrected P* using: (2040s) and the long-term (2090s) A1B scenario RCM ∗ b P = a ∗ P (3) data. Below is tabular summary for “a” and “b” values developed for each month (Table 2). Te impact of sampling variability is reduced by deter- mining the parameters a and b for every month period Comparison between bias corrected and uncorrected RCM of the year, including data from all years available (Lean- output data der and Buishand 2007). Te determination of the b Comparison was made among bias corrected RCM, parameter is done iteratively. Here the “Goal Seek” func- uncorrected RCM and observed data to show the inher- tion in the Microsoft excel served the purpose. It was ited biases in RCM on one hand and to validate the efec- determined such that the coefcient of variation (CV) tiveness of the factors a and b on the other hand. Te of the corrected daily precipitation matches the CV of comparisons are shown below with illustrative fgures for the observed daily precipitation. With the determined precipitation (Fig. 3), for maximum temperature (Fig. 4) parameter “b”, the transformed daily precipitation values and for minimum temperature (Fig. 5). are calculated using: Furthermore, it can be observed from Fig. 3 that the P∗ = Pb RCM and hence GCM has mainly underestimated the (4) precipitation over the basin in all months except in the Ten the parameter “a” is determined such that the month of November. In the contrary, the regional climate mean of the transformed daily values corresponds with model has generally overestimated the maximum tem- the observed mean. At the end, each block of 30 days has perature in all the months (Fig. 4). Figure 5 shows that got its own “a” and “b” parameters, which are the same the model overestimated minimum temperature except for each year. for months of November, December, and January. Tese biases in the model are attributed to poor resolution to Temperature bias correction: linear transformation method address the local forcing. Te correction of temperature only involves shifting and scaling to adjust the mean and variance (Leander and Climate change scenario Buishand 2007). Hence, for correcting the daily tempera- In this study, the future climate change scenarios for the ture a linear transformation technique is applied. For the 2040s and the 2090s were assessed in comparison with basin, the corrected daily temperature T­ was obtained corr the base period of the 2000s for three climatic variables as: – temperature, precipitation, and evapo-transpiration as T = aT + b corr un (5) presented below:

where ­Tun is the uncorrected daily temperature for the Temperature scenario base period (2000s) of RCM data, and a and b are linear Both the projected maximum and minimum tempera- constants. Te values of a and b are determined using the tures have generally shown an increasing trend for A1B Muleta Environ Syst Res (2021) 10:24 Page 8 of 15

Table 2 Tabular values of correction factors (a and b) used for bias correction Block/month Rainfall Max. temperature Min. temperature Remark A b a b a B

1 0.8792 1.3772 1.2805 10.6803 0.8133 2.3622 Applied to 2040s & 2090s, too − 2 1.3542 1.3361 0.8748 0.9254 0.8411 1.5618 Applied to 2040s & 2090s, too 3 1.1954 1.0213 1.1135 5.7307 0.8486 1.5090 Applied to 2040s & 2090s, too − 4 1.2072 1.0392 0.8521 0.9649 1.1131 2.3814 Applied to 2040s & 2090s, too − 5 1.5849 0.8936 0.9328 0.6110 0.8698 1.3467 Applied to 2040s & 2090s, too − 6 1.7855 0.8545 0.9972 2.0092 1.6076 9.1064 Applied to 2040s & 2090s, too − − 7 1.4089 0.8847 0.9127 0.0313 2.2550 18.3513 Applied to 2040s & 2090s, too − 8 1.8901 0.7863 0.7631 3.4377 0.8719 1.0104 Applied to 2040s & 2090s, too 9 1.2814 0.9810 0.8754 1.0238 0.6479 4.2501 Applied to 2040s & 2090s, too 10 0.9410 1.1642 0.6818 5.6713 0.5023 6.3036 Applied to 2040s & 2090s, too 11 0.6113 1.2509 0.5312 9.9613 0.3738 7.9866 Applied to 2040s & 2090s, too 12 1.4676 1.3667 0.8765 0.9123 0.2509 9.2255 Applied to 2040s & 2090s, too

10 OBSRCM(uncorrected) RCM(corrected) 9 8 7 6 5 all, mm 4 Rainf 3 2 1 0 Jan Feb MarApr May Jun Jul AugSep OctNov Dec Fig. 3 Comparison of bias corrected and uncorrected RCM precipitation data on the monthly basis

30 OBSRCM(uncorrected)RCM(corrected)

25 C o , e

r 20 u t a r e

p 15 m e T .

x 10 a M 5

0 Jan Feb MarApr MayJun JulAug SepOct NovDec Fig. 4 Comparison of bias corrected and uncorrected RCM data of maximum temperature on the monthly basis Muleta Environ Syst Res (2021) 10:24 Page 9 of 15

15 OBSRCM(uncorrected) RCM(corrected)

14.5

C 14 o , e r

u 13.5 t a r

e 13 p m e

T 12.5 . n i

M 12

11.5

11 Jan Feb MarApr MayJun JulAug SepOct NovDec Fig. 5 Comparison of bias corrected and uncorrected RCM data of minimum temperature on the monthly basis

emission scenario in both future time of 2040s and 2090s was predicted for the 2090s A1B global emission scenario (Fig. 6). Future projected maximum temperatures show as can be observed from Fig. 11. large temperature fuxes in the month of January both Moreover, one can observe considerable seasonal vari- for 2040s and 2090s forecasts with a temperature rise of ation in precipitation change with increasing trend in the 2.5 °C and 5.1 °C respectively as compared to the base rainy season (July to December) and a decreasing trend period. On the other hand, minimum temperature fux in the dry season (January to May), Fig. 11. Tis has an will be higher in the month of July for both mid-term and implication climate change on the intensifcation of long-term forecasts with a temperature rise of 2.5 °C and hydrologic extremes-drought and fooding in the future. 5.8 °C respectively (Fig. 7). Generally, the projected mean annual temperature increases from the baseline period by Evapo‑transpiration scenario an amount of 1 °C and 3.5 °C respectively, in 2040s and Relative to the current condition, the average annual 2090s. Tis implies that temperature will continue to evapo-transpiration over the basin shows increasing increase in the twenty frst century over the basin. trends in both short-term (2040s) and long-term (2090s) forecasts for the A1B scenario (Fig. 12). According to this Precipitation scenario study, evapotranspiration has been found to increase to In contrast to temperature, projection of rainfall does not an extent of 25% in the month of July (Fig. 13). Seasonal manifest a systematic increase or decrease in the coming wise, it can be observed that considerable evapotran- time horizon for A1B global emission scenario (Figs. 8, spiration fuxes are projected in summer season for the 9, 10). It shows a decreasing trend in the mid-term pro- 2090s and relatively little monthly and seasonal variation jection (Fig. 9) and an increasing trend for 2090s pro- is depicted in evapotranspiration of 2040s prediction. jection (Fig. 10). Tis contrasting trend implies the fact Te result shown is so logical in the sense that evapotran- that rainfall probability and frequency over a given area spiration will be favored by increased water surface cov- will be shifted spatially and temporally following change erage (in summer) and by the increasing temperature to in the climate. Precipitation experiences a mean annual the end of the century. decrease of 1.8% by 2040s and an increase of 1.8% in 2090s over the basin for the A1B emission scenario. Discussion However, future projection of rainfall depicts a con- Te current issue of climate change adaptive strat- siderable change on seasonal and monthly basis. For egy in the water resources sector can be achieved instance, the rainfall will experience a reduction of up to through the characterization of climate-related risks 29% in January and rises to 42% in February for future and response options to climate change under diferent 2040s projection. Likewise, a precipitation decreases of socio-economic futures and development prospects are up to 24% in January and a rise of up to 47% in December essential for sustainable water resources management Muleta Environ Syst Res (2021) 10:24 Page 10 of 15

Tmax Tmin 12000

10000 C

8000

ature, 6000

mper 4000 Te

2000

0

Year Fig. 6 Projected maximum and minimum annual temperature trends

Tmax-2040s Tmin-2040s Tmax-2090sTmin-2090s

6

5 C) ( 4

3

2 Temperature 1

0 Jan Feb MarApr May Jun JulAug Sep OctNov Dec Months Fig. 7 Future mean monthly max. and min. temperature fuxes compared to base period

PRECIP(2001-2100) Linear (PRECIP(2001-2100)) 3000

2500

2000 Precipitaon, mm 1500

1000

year Fig. 8 Annual precipitation trend in general (2001 through 2100) for A1B scenario Muleta Environ Syst Res (2021) 10:24 Page 11 of 15

PRECIP(2041-2050) Linear (PRECIP(2041-2050)) 2500

2000

1500 Precipitaon, mm

1000

year Fig. 9 Projected annual precipitation trend in the mid-term (2041–2050) for A1B scenario

PRECIP(2091-2100) Linear (PRECIP(2091-2100)) 3000

2500

2000 Precipitaon, mm

1500

1000

year Fig. 10 Projected annual precipitation trend in the long-term (2091–2100) for A1B scenario

(Cradock-Henry et al. 2018). Tis can be achieved does not manifest a general trend in the future scenarios; through relevant climate change scenarios development rather it depicts a slightly decreasing trend in the mid- at the national and local scales for sustainable develop- term forecast (2040s) and an increasing trend in the late ment and management of projects and programs across century (2090s) for the A1B emission scenario. Typically, relevant sectors (Aparecido et al. 2020; Poonia and Rao projection of rainfall has shown a large seasonal varia- 2018; Aziz et al. 2020). Te current study has laid a base- tion—a considerable increase in summer and autumn line information for climate change adaptive strategies with reduction in winter and spring (Fig. 11). Tis vari- and water resources management decision support. ation accompanied by highly increasing evapotranspira- Te projected maximum and minimum temperature tion in the future climate will have a clear implication of shows an increasing trend for the next century, which is a likely occurrence of hydrologic extremes-drought and obviously attributed to the increasing emission scenario fooding-over the basin. of GHG. But, on the contrary, Precipitation prediction Muleta Environ Syst Res (2021) 10:24 Page 12 of 15

2040s 2090s

60 50 40

%) 30

n( 20 10 0

Precipitao -10 -20 -30 -40 Jan Feb MarApr MayJun JulAug Sep OctNov Dec Months Fig. 11 Percentage monthly projected precipitation changes compared to the base period

ETo(2001-2100) Linear (ETo(2001-2100)) 1500.0

1400.0

1300.0

1200.0 ETo, mm/day

1100.0

1000.0

Year Fig. 12 Mean annual evapotranspiration trend for A1B emission scenario

Te results shown related to bias correction indicates biases in the model outputs are of course expected owed that still considerable biases inherited from GCM are to locally sensitive climatic variables and complexity of prevailing in the RCM outputs. Studies have shown that the global system, in which case GCM and even RCM are Global climate models (GCMs) are inherently unable to inherently unable to represent local sub grid-scale fea- present local subgrid-scale features and dynamics (David tures and dynamics at basin levels. and José 2014) and consequently climate change studies Te current study outputs of climate change on the at local level are required for decision makers. Bastien basin shows a good agreement with previous studies (2017). Furthermore, it is an indication of a poor rep- such as Carvalho et al. (2017), Costa et al. (2019), Hany resentation of the regional or continental level climate et al. (2016), Myo and Zin (2020), Roth et al. (2018), and change scenario at the basin level. Tese variations and Xueping et al. (2018) (Table 1) as far as temperature Muleta Environ Syst Res (2021) 10:24 Page 13 of 15

2040s 2090s

30

25 %)

n( 20

15

10

5 Evapotranspirao 0

-5 Jan Feb MarApr MayJun JulAug Sep OctNov Dec Months Fig. 13 Percentage changes in the projected evapotranspiration as compared to the base period

change scenario is considered. Moreover, the current research needs also to be extended to the other basins of projected mean annual temperature changes of 3.5 °C in the country. 2090s is within the range projected by (IPCC 2018) which indicated that the average temperature rise to be in the Conclusions range of 1.4–5.8 °C towards the end of this century. But in the case of precipitation change scenario, the results Te result of climate projection reveals signifcant climate do not show consistency and of course this is true also change scenarios over Baro-Akobo basin particularly with the previous studies and we can say that the result in terms of monthly and seasonal distribution with the showed a good agreement with the studies in behavior implication of hydrologic extremes. Tis, accompanied rather than quantitatively (refer to Table 1). with the extensive water resources development projects Te inconsistency in the rainfall change scenario in the basin and its transboundary environmental efect, among the studies is attributed to the sensitivity of rain- calls for climate change adaptive strategies and real-time fall variable to the poor resolutions of GCM (Myo and water resources managements. Terefore, the obtained Zin 2020), diferences in the models (Costa et al. 2019), result produced a baseline data that support adaptation and local factors such as orographic, coastal and veg- of the vast water resources development planning and etation efects (Jose et al. 2016; David and José 2014). Of management in the basin to the ever-changing climate. course, this is the central reason to have conducted the Te author highly recommends researchers to model the current study in addition to the peculiar features of the relation between these climate change scenarios and res- basin such as distinctive bimodal rainfall pattern and ervoir infows as climate change adaptation strategy. It the vast water resources developments. To this end, bias further helps reservoir operators to modify their opera- correction is employed in the current study (which, of tion rules to better combat hydrologic extremes and course, lacks in the previous studies) to better account minimize the possible environmental impacts of the vast for the local efects than directly using the downscaled water resources developments on the local and down- meteorological data as in the case of previous studies. stream Nile basin countries. However, the current study is limited to a single alter- However, the study has involved several models’ out- native scenario-A1B, with the objective of getting single puts where each possessed a certain level of uncertainty. valued average climate scenarios. In fact, the probabil- Moreover, it is based on a single emission scenario sto- ity of a single-story line to occur in the future is highly ryline (A1B). Hence, the results of this study should limited. Terefore, further studies can be extended using be taken with care and be considered as indicative of the other alternative scenarios and models to show the the likely future rather than accurate predictions. It is range of climate change scenarios over the basin. Te believed that the results of this study give a baseline data Muleta Environ Syst Res (2021) 10:24 Page 14 of 15

and increase awareness on the possible future risks of cli- Aparecido LEO, José RSCM, Kamila CM, Pedro AL, João AL, Gabriel HOS, Guil- herme BT (2020) Agricultural zoning as tool for expansion of cassava in mate change. climate change scenarios. Theor Appl Climatol. 142(3):1085–1095 Aziz R, Yucel I, Yozgatligil C (2020) Nonstationarity impacts on frequency analysis of yearly and seasonal extreme temperature in Turkey. Atmos Res Abbreviations 238:104875 AOGCMs: Atmosphere–Ocean General Circulation Models; CV: Coefcient of Barrow E, Hulme M, Semenov M (1996) Efect of using diferent methods in Variation; DEM: Digital Elevation Model; ETo: Reference Evapotranspiration; the construction of climate change scenarios: examples from Europe. FAO: Food and Agricultural Organization; GCM: Global Climate Model; GCP: Climate Res 7(3):195–211 Grid Center Point; GHG: Greenhouse Gas; HEC-GeoHMS: Hydrologic Engineer- Bastien B (2017) Downscaling climate change signal in Mexico ing Center Geospatial Hydrologic Modeling System; IWMI: International Water Bell WP, Wild P, Froome C, Wagner LD (2013) Selecting emission and climate Management Institute; IPCC: Intergovernmental Panel for Climate Change; change scenarios: review. University of Queensland, Queensland LAMs: Limited area models; MAP: Mean Areal Precipitation; MoWR: Ministry Brekke LD, Maurer EP, Anderson JD, Dettinger MD, Townsely ES, Harrison A, of Water Resources; NMA: National Meteorological Agency; NAPA: National Pruitt T (2009) Assessing reservoir operations risk under climate change. Adaptation Programme of Action; Obs: Observed; SRES: Special Report on Water Resour Res 45:7 Emission Scenario; RCM: Regional Climate Model; WMO: World Meteorological Carvalho SC, Santos FD, Pulquério M (2017) Climate change scenarios for Organization. Angola: an analysis of precipitation and temperature projections using four RCMs. Int J Climatol 37(8):3398–3412 Acknowledgements Chow V, Maidment D, Mays L (1988) Applied Hydrology then. McGraw-Hill It is my pleasure to express my special heartfelt thanks and appreciation to Book Company, New York my advisor Dr. Knolmar Marcell, an Assistant Professor at the Department of Conde C, Estrada F, Martínez B, Sánchez O, Gay C (2011) Regional climate Sanitary and Environmental Engineering, Budapest University of Technology change scenarios for México. Atmósfera 24(1):125–140 and Economics, Budapest, Hungary for his valuable support in draft review Conway D, Hulme M (1996) The impacts of climate variability and future cli- and editing of the manuscript and his all-round advisory support. I would mate change in the Nile Basin on Water Resources in Egypt. Water Resour like to extend my gratitude to Dr. Solomon Seyoum, a director general for Develop 12(3):277–296 International Water Management Institute at Ethiopia for his excellent support Costa RL, Heliofábio BG, Fabrício DSS, Rodrigo LRJ (2019) Downscale of future in providing GCM data and his guidance on how to bias correct the data. I climate change scenarios applied to Recife-PE. J Hyperspectral Rem Sens. am also highly indebted to all organizations and Institutes those provided me https://​doi.​org/​10.​1007/​s00704-​020-​03367-1 with important data & materials mainly Ministry of Water and Energy, National Cradock-Henry NA, Frame B, Preston BL, Reisinger A, Rothman DS (2018) Metrological Agency, and International Water Management Institute. Dynamic adaptive pathways in downscaled climate change scenarios. Clim Change 150(3):333–341 Authors’ contributions Cubash U, von Storch H, Waszkewitz J, Zorita E (1996) Estimates of climate TN performed all the conceptualization, study design, statistical analysis of change in Southern Europe derived from dynamical climate model results, data interpretation, and writing the manuscript. The author read and output. Climate Res 7:129–149 approved the fnal manuscript. David M, José MA (2014) Neural networks for downscaling climate change scenarios Funding Raje D, Mujumdar PP (2010) Reservoir performance under uncertainty in This study was not funded by any grant. hydrologic impacts of climate change. Adv Water Resour 33(3):312–326 Dickinson RE, Errico RM, Giorgi F, Bates GT (1989) A regional climate model for Availability of data and materials the western United States. Clim Change 15(3):383–422 The datasets used and/or analyzed during the current study are included as Dickinson RE, Bougher SW (1986) Venus mesosphere and thermosphere: 1. much as possible and available from the corresponding author up on request. Heat budget and thermal structure. J Geophys Res 91(A1):70–80 Feyissa G, Zeleke G, Bewket W, Gebremariam E (2018) Downscaling of future Declarations temperature and precipitation extremes in Addis Ababa under climate change. Climate 6(3):58 Mostafa H, Saleh H, El Sheikh M, Kheireldin K (2016) Assessing the impacts of Ethics approval and consent to participate climate changes on the eastern Nile fow at Aswan. J Am Sci 12(1):1–9 Not applicable. Hewitson BC, Crane RG (1996) Climate downscaling: techniques and applica- tion. Clim Res 7(2):85–95 Consent for publication Hulme M, Viner D (1998) A climate change scenario for the Tropics. Climate Not applicable. change (UNEP/IES), pp 145–176 Ines AV, Hansen JW (2006) Bias correction of daily GCM rainfall for crop simula- Competing interests tion studies. Agric For Meteorol 138(1–4):44–53 The author declares that there is not any competing interest. IPCC (2001) Climate Change 2001: Impacts, Adaptation and Vulnerability Contribution of Working Group II to the Third Assessment Report of the Received: 13 November 2020 Accepted: 22 February 2021 IPCC. Cambridge, Cambridge University Press IPCC (2000) Emission Scenarios: A special Report of Working Group III of the IPCC. Cambridge University Press, Cambridge IPCC (2007) Impacts, Adaptation and Vulnerability: Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel References on Climate Change. Cambridge University Press, Cambridge Alemayehu T, Kebede S, Liu L, Nedaw D (2016) Groundwater recharge under IPCC (2018) Special Report on the impacts of global warming of 1.5°C above changing landuses and climate variability: the case of Baro-Akobo River pre-industrial levels and related global greenhouse gas emission path- Basin, Ethiopia. J Environ Earth Sci 6(1):78–95 ways. Switzerland Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration-Guide- Jones JA (1999) Climate change and sustainable water resources: placing the lines for computing crop water requirements-FAO Irrigation and drainage threat of global warming in perspective. Hydrol Sci J 44(4):541–557 paper 56. Fao, Rome 300(9):D05109 Jose AM, Alves LM, Torres RR (2016) Regional climate change scenarios in the Anthes RA, Kuo YH, Baumhefner DP, Errico RM, Bettge TW (1985) Predictability Brazilian Pantanal watershed. Climate Res 68(2–3):201–213 of mesoscale atmospheric motions. Adv Geophys 28:159–202 Muleta Environ Syst Res (2021) 10:24 Page 15 of 15

Leander R, Buishand TA (2007) Resampling of regional climate model output Santosh N, Pradhananga S, Lutz A, Shrestha A, Arun SB (2019) Climate Change for the simulation of extreme river fows. J Hydrol 332(3–4):487–496 Scenarios for Nepal for National Adaptation Plan. Ministry of Forests and Li L, Xu H, Chen X, Simonovic SP (2010) Streamfow forecast and reservoir Environment operation performance assessment under climate change. Water Resour Schneider SH, Easterling WE, Mearns LO (2000) Adaptation: sensitivity to Manage 24(1):83–104 natural variability, agent assumptions and dynamic climate changes. Clim De Luca DL, Petroselli A, Galasso L (2020) A transient stochastic rainfall genera- Change 2000(45):203–221 tor for climate changes analysis at hydrological scales in Central Italy. Smith JB, Pitts G (1997) Regional Climate change scenariosfor vulnerability and Atmosphere 11(12):1292 adaptation assessments. Clim Change 36:3–21 Mearns LO, Bogardi I, Giorgi F, Matyasovzky I, Palecki M (1999) Comparison of Tahani MS, ElShamy M, Mohammed AA, Abbas MS (2013) The Development climate change scenarios generated from regional climate model experi- of the Baro-Akobo-Sobat sub basin and its Impact on Downstream Nile ments and statistical downscaling. J Geophys Res 104(D6):6603–6621 Basin Countries. Nile Water Sci Eng J. 6:2 Mohamed H (2015) Impact of Climate Change on Water Resources in the Nile Wilby RL, Hay LE, Gutowski WJ Jr, Arritt RW, Takle ES, Pan Z, Leavesley GH, Clark Basin and Adaptation measures in Egypt. Researchgate. https://​www.​ MP (2000) Hydrological responses to dynamically and statistically down- resea​rchga​te.​net/​publi​cation/​27733​0599. Accessed 25 July 2020 scaled climate model output. Geophys Res Lett 62:182 Myo HT, Zin WW (2020) Forecasting climate change scenarios in central dry Wilby RL, Wigley TML (1997) Downscaling general circulation model output: a zone of Myanmar review of methods and limitations. Prog Phys Geogr 21(4):530–548 NAPA (2007) Preparation of National Adaptation Programme of Action (NAPA) Xueping Z, Chi Z, Qi W, Wenjun C, Xuehua Z, Xueni W (2018) Multiple climate of Ethiopia. Addis Abeba change scenarios and runof response in Biliu River. Water 10:126 Pabitra G, Praju G (2020) Climate Change Scenarios in Nepal. Technical, Kathmandu Poonia S, Rao AS (2018) Climate and climate change scenarios in the Indian Publisher’s Note Thar Region. Springer, Switzerland Springer Nature remains neutral with regard to jurisdictional claims in pub- Rahman HA (2018) Climate change scenarios in Malaysia: engaging the public. lished maps and institutional afliations. Int J Malay-Nusantara Stud 12:7 Robert K, Colin H (2007) A review of climate change and its potential impacts on water resources in the UK. Ofcial Publication of the European Water Association (EWA), London Roth V, Lemann T, Zeleke G, Subhatu AT (2018) Efects of climate change on water resources in the upper Blue Nile Basin of Ethiopia. Heliyon 4:e0071 Safeh J, Rebwar D, Forough J (2020) Climate change scenarios and efects on snow-melt runof. Civil Eng J. 6:1715